Monday, January 17, 2011

Computing for humanity in action: Ushahidi for the Australian floods

I gave a talk some weeks back on "computing for humanity", or how the collective social conscience of us computer scientists obliges us to explore ways in which our work can find immediate humanitarian application. I started the talk by discussing Ushahidi (Swahili for "testimony"), an open source application that was initially set up for crowdmapping violence in Kenya following its recent presidential elections, and subsequently used during the Haiti earthquakes and elsewhere (see www.ushahidi.com). Crowdmapping taps into the wisdom of the masses by generating an interactive spatio-temporal visualization of events on Open StreetMaps or Google Maps - these events being reported by citizen journalists via SMS (often the last remaining mobile service when all else fails in a disaster) or via email etc.

So my colleague and long-standing associate of the Decision Systems Lab, Lothar Hinsche, went and walked the talk. In the space of 11 hours last Tuesday, just as the flood waters hit Brisbane, he set up the Ushahidi platform for crowdmapping that disaster - see the site at brisbaneflood.au.ms

That was quite a feat. We tried hard to get an SMS gateway from major telcos, but that's been a challenge. We also tried to get some way of pubicizing the site so that more people might SMS event reports in, but that too has been a challenge. But the site has been up, bravely doing its thing, despite all of this...

In the meantime, the research dimension has been intriguing me for a while. Adaptive case management instead of BPM, driven by crowdsourced spatio-temporal event maps? Complex event processing? More to come...